1 /* $Id: mdoc_argv.c,v 1.109 2016/08/28 16:15:12 schwarze Exp $ */
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2012, 2014, 2015 Ingo Schwarze <schwarze@openbsd.org>
6 * Permission to use, copy, modify, and distribute this software for any
7 * purpose with or without fee is hereby granted, provided that the above
8 * copyright notice and this permission notice appear in all copies.
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
20 #include <sys/types.h>
27 #include "mandoc_aux.h"
31 #include "libmandoc.h"
35 #define MULTI_STEP 5 /* pre-allocate argument values */
36 #define DELIMSZ 6 /* max possible size of a delimiter */
40 ARGSFL_DELIM
, /* handle delimiters of [[::delim::][ ]+]+ */
41 ARGSFL_TABSEP
/* handle tab/`Ta' separated phrases */
45 ARGV_NONE
, /* no args to flag (e.g., -split) */
46 ARGV_SINGLE
, /* one arg to flag (e.g., -file xxx) */
47 ARGV_MULTI
/* multiple args (e.g., -column xxx yyy) */
52 const enum mdocargt
*argvs
;
55 static void argn_free(struct mdoc_arg
*, int);
56 static enum margserr
args(struct roff_man
*, int, int *,
57 char *, enum argsflag
, char **);
58 static int args_checkpunct(const char *, int);
59 static void argv_multi(struct roff_man
*, int,
60 struct mdoc_argv
*, int *, char *);
61 static void argv_single(struct roff_man
*, int,
62 struct mdoc_argv
*, int *, char *);
64 static const enum argvflag argvflags
[MDOC_ARG_MAX
] = {
65 ARGV_NONE
, /* MDOC_Split */
66 ARGV_NONE
, /* MDOC_Nosplit */
67 ARGV_NONE
, /* MDOC_Ragged */
68 ARGV_NONE
, /* MDOC_Unfilled */
69 ARGV_NONE
, /* MDOC_Literal */
70 ARGV_SINGLE
, /* MDOC_File */
71 ARGV_SINGLE
, /* MDOC_Offset */
72 ARGV_NONE
, /* MDOC_Bullet */
73 ARGV_NONE
, /* MDOC_Dash */
74 ARGV_NONE
, /* MDOC_Hyphen */
75 ARGV_NONE
, /* MDOC_Item */
76 ARGV_NONE
, /* MDOC_Enum */
77 ARGV_NONE
, /* MDOC_Tag */
78 ARGV_NONE
, /* MDOC_Diag */
79 ARGV_NONE
, /* MDOC_Hang */
80 ARGV_NONE
, /* MDOC_Ohang */
81 ARGV_NONE
, /* MDOC_Inset */
82 ARGV_MULTI
, /* MDOC_Column */
83 ARGV_SINGLE
, /* MDOC_Width */
84 ARGV_NONE
, /* MDOC_Compact */
85 ARGV_NONE
, /* MDOC_Std */
86 ARGV_NONE
, /* MDOC_Filled */
87 ARGV_NONE
, /* MDOC_Words */
88 ARGV_NONE
, /* MDOC_Emphasis */
89 ARGV_NONE
, /* MDOC_Symbolic */
90 ARGV_NONE
/* MDOC_Symbolic */
93 static const enum mdocargt args_Ex
[] = {
98 static const enum mdocargt args_An
[] = {
104 static const enum mdocargt args_Bd
[] = {
116 static const enum mdocargt args_Bf
[] = {
123 static const enum mdocargt args_Bk
[] = {
128 static const enum mdocargt args_Bl
[] = {
147 static const struct mdocarg mdocargs
[MDOC_MAX
] = {
148 { ARGSFL_DELIM
, NULL
}, /* Ap */
149 { ARGSFL_NONE
, NULL
}, /* Dd */
150 { ARGSFL_NONE
, NULL
}, /* Dt */
151 { ARGSFL_NONE
, NULL
}, /* Os */
152 { ARGSFL_NONE
, NULL
}, /* Sh */
153 { ARGSFL_NONE
, NULL
}, /* Ss */
154 { ARGSFL_NONE
, NULL
}, /* Pp */
155 { ARGSFL_DELIM
, NULL
}, /* D1 */
156 { ARGSFL_DELIM
, NULL
}, /* Dl */
157 { ARGSFL_NONE
, args_Bd
}, /* Bd */
158 { ARGSFL_NONE
, NULL
}, /* Ed */
159 { ARGSFL_NONE
, args_Bl
}, /* Bl */
160 { ARGSFL_NONE
, NULL
}, /* El */
161 { ARGSFL_NONE
, NULL
}, /* It */
162 { ARGSFL_DELIM
, NULL
}, /* Ad */
163 { ARGSFL_DELIM
, args_An
}, /* An */
164 { ARGSFL_DELIM
, NULL
}, /* Ar */
165 { ARGSFL_DELIM
, NULL
}, /* Cd */
166 { ARGSFL_DELIM
, NULL
}, /* Cm */
167 { ARGSFL_DELIM
, NULL
}, /* Dv */
168 { ARGSFL_DELIM
, NULL
}, /* Er */
169 { ARGSFL_DELIM
, NULL
}, /* Ev */
170 { ARGSFL_NONE
, args_Ex
}, /* Ex */
171 { ARGSFL_DELIM
, NULL
}, /* Fa */
172 { ARGSFL_NONE
, NULL
}, /* Fd */
173 { ARGSFL_DELIM
, NULL
}, /* Fl */
174 { ARGSFL_DELIM
, NULL
}, /* Fn */
175 { ARGSFL_DELIM
, NULL
}, /* Ft */
176 { ARGSFL_DELIM
, NULL
}, /* Ic */
177 { ARGSFL_DELIM
, NULL
}, /* In */
178 { ARGSFL_DELIM
, NULL
}, /* Li */
179 { ARGSFL_NONE
, NULL
}, /* Nd */
180 { ARGSFL_DELIM
, NULL
}, /* Nm */
181 { ARGSFL_DELIM
, NULL
}, /* Op */
182 { ARGSFL_DELIM
, NULL
}, /* Ot */
183 { ARGSFL_DELIM
, NULL
}, /* Pa */
184 { ARGSFL_NONE
, args_Ex
}, /* Rv */
185 { ARGSFL_DELIM
, NULL
}, /* St */
186 { ARGSFL_DELIM
, NULL
}, /* Va */
187 { ARGSFL_DELIM
, NULL
}, /* Vt */
188 { ARGSFL_DELIM
, NULL
}, /* Xr */
189 { ARGSFL_NONE
, NULL
}, /* %A */
190 { ARGSFL_NONE
, NULL
}, /* %B */
191 { ARGSFL_NONE
, NULL
}, /* %D */
192 { ARGSFL_NONE
, NULL
}, /* %I */
193 { ARGSFL_NONE
, NULL
}, /* %J */
194 { ARGSFL_NONE
, NULL
}, /* %N */
195 { ARGSFL_NONE
, NULL
}, /* %O */
196 { ARGSFL_NONE
, NULL
}, /* %P */
197 { ARGSFL_NONE
, NULL
}, /* %R */
198 { ARGSFL_NONE
, NULL
}, /* %T */
199 { ARGSFL_NONE
, NULL
}, /* %V */
200 { ARGSFL_DELIM
, NULL
}, /* Ac */
201 { ARGSFL_NONE
, NULL
}, /* Ao */
202 { ARGSFL_DELIM
, NULL
}, /* Aq */
203 { ARGSFL_DELIM
, NULL
}, /* At */
204 { ARGSFL_DELIM
, NULL
}, /* Bc */
205 { ARGSFL_NONE
, args_Bf
}, /* Bf */
206 { ARGSFL_NONE
, NULL
}, /* Bo */
207 { ARGSFL_DELIM
, NULL
}, /* Bq */
208 { ARGSFL_DELIM
, NULL
}, /* Bsx */
209 { ARGSFL_DELIM
, NULL
}, /* Bx */
210 { ARGSFL_NONE
, NULL
}, /* Db */
211 { ARGSFL_DELIM
, NULL
}, /* Dc */
212 { ARGSFL_NONE
, NULL
}, /* Do */
213 { ARGSFL_DELIM
, NULL
}, /* Dq */
214 { ARGSFL_DELIM
, NULL
}, /* Ec */
215 { ARGSFL_NONE
, NULL
}, /* Ef */
216 { ARGSFL_DELIM
, NULL
}, /* Em */
217 { ARGSFL_NONE
, NULL
}, /* Eo */
218 { ARGSFL_DELIM
, NULL
}, /* Fx */
219 { ARGSFL_DELIM
, NULL
}, /* Ms */
220 { ARGSFL_DELIM
, NULL
}, /* No */
221 { ARGSFL_DELIM
, NULL
}, /* Ns */
222 { ARGSFL_DELIM
, NULL
}, /* Nx */
223 { ARGSFL_DELIM
, NULL
}, /* Ox */
224 { ARGSFL_DELIM
, NULL
}, /* Pc */
225 { ARGSFL_DELIM
, NULL
}, /* Pf */
226 { ARGSFL_NONE
, NULL
}, /* Po */
227 { ARGSFL_DELIM
, NULL
}, /* Pq */
228 { ARGSFL_DELIM
, NULL
}, /* Qc */
229 { ARGSFL_DELIM
, NULL
}, /* Ql */
230 { ARGSFL_NONE
, NULL
}, /* Qo */
231 { ARGSFL_DELIM
, NULL
}, /* Qq */
232 { ARGSFL_NONE
, NULL
}, /* Re */
233 { ARGSFL_NONE
, NULL
}, /* Rs */
234 { ARGSFL_DELIM
, NULL
}, /* Sc */
235 { ARGSFL_NONE
, NULL
}, /* So */
236 { ARGSFL_DELIM
, NULL
}, /* Sq */
237 { ARGSFL_NONE
, NULL
}, /* Sm */
238 { ARGSFL_DELIM
, NULL
}, /* Sx */
239 { ARGSFL_DELIM
, NULL
}, /* Sy */
240 { ARGSFL_DELIM
, NULL
}, /* Tn */
241 { ARGSFL_DELIM
, NULL
}, /* Ux */
242 { ARGSFL_DELIM
, NULL
}, /* Xc */
243 { ARGSFL_NONE
, NULL
}, /* Xo */
244 { ARGSFL_NONE
, NULL
}, /* Fo */
245 { ARGSFL_DELIM
, NULL
}, /* Fc */
246 { ARGSFL_NONE
, NULL
}, /* Oo */
247 { ARGSFL_DELIM
, NULL
}, /* Oc */
248 { ARGSFL_NONE
, args_Bk
}, /* Bk */
249 { ARGSFL_NONE
, NULL
}, /* Ek */
250 { ARGSFL_NONE
, NULL
}, /* Bt */
251 { ARGSFL_NONE
, NULL
}, /* Hf */
252 { ARGSFL_DELIM
, NULL
}, /* Fr */
253 { ARGSFL_NONE
, NULL
}, /* Ud */
254 { ARGSFL_DELIM
, NULL
}, /* Lb */
255 { ARGSFL_NONE
, NULL
}, /* Lp */
256 { ARGSFL_DELIM
, NULL
}, /* Lk */
257 { ARGSFL_DELIM
, NULL
}, /* Mt */
258 { ARGSFL_DELIM
, NULL
}, /* Brq */
259 { ARGSFL_NONE
, NULL
}, /* Bro */
260 { ARGSFL_DELIM
, NULL
}, /* Brc */
261 { ARGSFL_NONE
, NULL
}, /* %C */
262 { ARGSFL_NONE
, NULL
}, /* Es */
263 { ARGSFL_DELIM
, NULL
}, /* En */
264 { ARGSFL_DELIM
, NULL
}, /* Dx */
265 { ARGSFL_NONE
, NULL
}, /* %Q */
266 { ARGSFL_NONE
, NULL
}, /* br */
267 { ARGSFL_NONE
, NULL
}, /* sp */
268 { ARGSFL_NONE
, NULL
}, /* %U */
269 { ARGSFL_NONE
, NULL
}, /* Ta */
270 { ARGSFL_NONE
, NULL
}, /* ll */
275 * Parse flags and their arguments from the input line.
276 * These come in the form -flag [argument ...].
277 * Some flags take no argument, some one, some multiple.
280 mdoc_argv(struct roff_man
*mdoc
, int line
, int tok
,
281 struct mdoc_arg
**reta
, int *pos
, char *buf
)
283 struct mdoc_argv tmpv
;
284 struct mdoc_argv
**retv
;
285 const enum mdocargt
*argtable
;
292 /* Which flags does this macro support? */
294 argtable
= mdocargs
[tok
].argvs
;
295 if (argtable
== NULL
)
298 /* Loop over the flags on the input line. */
301 while (buf
[ipos
] == '-') {
303 /* Seek to the first unescaped space. */
305 for (argname
= buf
+ ++ipos
; buf
[ipos
] != '\0'; ipos
++)
306 if (buf
[ipos
] == ' ' && buf
[ipos
- 1] != '\\')
310 * We want to nil-terminate the word to look it up.
311 * But we may not have a flag, in which case we need
312 * to restore the line as-is. So keep around the
313 * stray byte, which we'll reset upon exiting.
316 if ((savechar
= buf
[ipos
]) != '\0')
320 * Now look up the word as a flag. Use temporary
321 * storage that we'll copy into the node's flags.
324 while ((tmpv
.arg
= *argtable
++) != MDOC_ARG_MAX
)
325 if ( ! strcmp(argname
, mdoc_argnames
[tmpv
.arg
]))
328 /* If it isn't a flag, restore the saved byte. */
330 if (tmpv
.arg
== MDOC_ARG_MAX
) {
331 if (savechar
!= '\0')
332 buf
[ipos
- 1] = savechar
;
336 /* Read to the next word (the first argument). */
338 while (buf
[ipos
] == ' ')
341 /* Parse the arguments of the flag. */
348 switch (argvflags
[tmpv
.arg
]) {
350 argv_single(mdoc
, line
, &tmpv
, &ipos
, buf
);
353 argv_multi(mdoc
, line
, &tmpv
, &ipos
, buf
);
359 /* Append to the return values. */
362 *reta
= mandoc_calloc(1, sizeof(**reta
));
364 retc
= ++(*reta
)->argc
;
365 retv
= &(*reta
)->argv
;
366 *retv
= mandoc_reallocarray(*retv
, retc
, sizeof(**retv
));
367 memcpy(*retv
+ retc
- 1, &tmpv
, sizeof(**retv
));
369 /* Prepare for parsing the next flag. */
372 argtable
= mdocargs
[tok
].argvs
;
377 mdoc_argv_free(struct mdoc_arg
*p
)
391 for (i
= (int)p
->argc
- 1; i
>= 0; i
--)
399 argn_free(struct mdoc_arg
*p
, int iarg
)
401 struct mdoc_argv
*arg
;
404 arg
= &p
->argv
[iarg
];
406 if (arg
->sz
&& arg
->value
) {
407 for (j
= (int)arg
->sz
- 1; j
>= 0; j
--)
412 for (--p
->argc
; iarg
< (int)p
->argc
; iarg
++)
413 p
->argv
[iarg
] = p
->argv
[iarg
+1];
417 mdoc_args(struct roff_man
*mdoc
, int line
, int *pos
,
418 char *buf
, int tok
, char **v
)
426 fl
= tok
== TOKEN_NONE
? ARGSFL_NONE
: mdocargs
[tok
].flags
;
428 return args(mdoc
, line
, pos
, buf
, fl
, v
);
431 * We know that we're in an `It', so it's reasonable to expect
432 * us to be sitting in a `Bl'. Someday this may not be the case
433 * (if we allow random `It's sitting out there), so provide a
434 * safe fall-back into the default behaviour.
437 for (n
= mdoc
->last
; n
; n
= n
->parent
)
438 if (MDOC_Bl
== n
->tok
)
439 if (LIST_column
== n
->norm
->Bl
.type
) {
444 return args(mdoc
, line
, pos
, buf
, fl
, v
);
448 args(struct roff_man
*mdoc
, int line
, int *pos
,
449 char *buf
, enum argsflag fl
, char **v
)
454 if (buf
[*pos
] == '\0') {
455 if (mdoc
->flags
& MDOC_PHRASELIT
&&
456 ! (mdoc
->flags
& MDOC_PHRASE
)) {
457 mandoc_msg(MANDOCERR_ARG_QUOTE
,
458 mdoc
->parse
, line
, *pos
, NULL
);
459 mdoc
->flags
&= ~MDOC_PHRASELIT
;
466 if (fl
== ARGSFL_DELIM
&& args_checkpunct(buf
, *pos
))
470 * Tabs in `It' lines in `Bl -column' can't be escaped.
471 * Phrases are reparsed for `Ta' and other macros later.
474 if (fl
== ARGSFL_TABSEP
) {
475 if ((p
= strchr(*v
, '\t')) != NULL
) {
478 * Words right before and right after
479 * tab characters are not parsed,
480 * unless there is a blank in between.
483 if (p
> buf
&& p
[-1] != ' ')
484 mdoc
->flags
|= MDOC_PHRASEQL
;
486 mdoc
->flags
|= MDOC_PHRASEQN
;
489 * One or more blanks after a tab cause
490 * one leading blank in the next column.
491 * So skip all but one of them.
494 *pos
+= (int)(p
- *v
) + 1;
495 while (buf
[*pos
] == ' ' && buf
[*pos
+ 1] == ' ')
499 * A tab at the end of an input line
500 * switches to the next column.
503 if (buf
[*pos
] == '\0' || buf
[*pos
+ 1] == '\0')
504 mdoc
->flags
|= MDOC_PHRASEQN
;
506 p
= strchr(*v
, '\0');
508 mandoc_msg(MANDOCERR_SPACE_EOL
,
509 mdoc
->parse
, line
, *pos
, NULL
);
510 *pos
+= (int)(p
- *v
);
513 /* Skip any trailing blank characters. */
514 while (p
> *v
&& p
[-1] == ' ' &&
515 (p
- 1 == *v
|| p
[-2] != '\\'))
523 * Process a quoted literal. A quote begins with a double-quote
524 * and ends with a double-quote NOT preceded by a double-quote.
525 * NUL-terminate the literal in place.
526 * Collapse pairs of quotes inside quoted literals.
527 * Whitespace is NOT involved in literal termination.
530 if (mdoc
->flags
& MDOC_PHRASELIT
|| buf
[*pos
] == '\"') {
531 if ( ! (mdoc
->flags
& MDOC_PHRASELIT
))
534 if (mdoc
->flags
& MDOC_PHRASE
)
535 mdoc
->flags
|= MDOC_PHRASELIT
;
538 for ( ; buf
[*pos
]; (*pos
)++) {
539 /* Move following text left after quoted quotes. */
541 buf
[*pos
- pairs
] = buf
[*pos
];
542 if ('\"' != buf
[*pos
])
544 /* Unquoted quotes end quoted args. */
545 if ('\"' != buf
[*pos
+ 1])
547 /* Quoted quotes collapse. */
552 buf
[*pos
- pairs
] = '\0';
554 if (buf
[*pos
] == '\0') {
555 if ( ! (mdoc
->flags
& MDOC_PHRASE
))
556 mandoc_msg(MANDOCERR_ARG_QUOTE
,
557 mdoc
->parse
, line
, *pos
, NULL
);
561 mdoc
->flags
&= ~MDOC_PHRASELIT
;
562 buf
[(*pos
)++] = '\0';
564 if ('\0' == buf
[*pos
])
567 while (' ' == buf
[*pos
])
570 if ('\0' == buf
[*pos
])
571 mandoc_msg(MANDOCERR_SPACE_EOL
, mdoc
->parse
,
578 *v
= mandoc_getarg(mdoc
->parse
, &p
, line
, pos
);
581 * After parsing the last word in this phrase,
582 * tell lookup() whether or not to interpret it.
585 if (*p
== '\0' && mdoc
->flags
& MDOC_PHRASEQL
) {
586 mdoc
->flags
&= ~MDOC_PHRASEQL
;
587 mdoc
->flags
|= MDOC_PHRASEQF
;
593 * Check if the string consists only of space-separated closing
594 * delimiters. This is a bit of a dance: the first must be a close
595 * delimiter, but it may be followed by middle delimiters. Arbitrary
596 * whitespace may separate these tokens.
599 args_checkpunct(const char *buf
, int i
)
605 /* First token must be a close-delimiter. */
607 for (j
= 0; buf
[i
] && ' ' != buf
[i
] && j
< DELIMSZ
; j
++, i
++)
614 if (DELIM_CLOSE
!= mdoc_isdelim(dbuf
))
617 while (' ' == buf
[i
])
620 /* Remaining must NOT be open/none. */
624 while (buf
[i
] && ' ' != buf
[i
] && j
< DELIMSZ
)
625 dbuf
[j
++] = buf
[i
++];
631 d
= mdoc_isdelim(dbuf
);
632 if (DELIM_NONE
== d
|| DELIM_OPEN
== d
)
635 while (' ' == buf
[i
])
639 return '\0' == buf
[i
];
643 argv_multi(struct roff_man
*mdoc
, int line
,
644 struct mdoc_argv
*v
, int *pos
, char *buf
)
649 for (v
->sz
= 0; ; v
->sz
++) {
650 if (buf
[*pos
] == '-')
652 ac
= args(mdoc
, line
, pos
, buf
, ARGSFL_NONE
, &p
);
656 if (v
->sz
% MULTI_STEP
== 0)
657 v
->value
= mandoc_reallocarray(v
->value
,
658 v
->sz
+ MULTI_STEP
, sizeof(char *));
660 v
->value
[(int)v
->sz
] = mandoc_strdup(p
);
665 argv_single(struct roff_man
*mdoc
, int line
,
666 struct mdoc_argv
*v
, int *pos
, char *buf
)
671 ac
= args(mdoc
, line
, pos
, buf
, ARGSFL_NONE
, &p
);
676 v
->value
= mandoc_malloc(sizeof(char *));
677 v
->value
[0] = mandoc_strdup(p
);